bigNOSTRals

The open-source analytics & data-modeling layer for Nostr.

A dbt project that turns raw relay events into clean, queryable tables. Researchers, journalists, and curious engineers can fork it, run it, and smell what the network is really saying.

"The network has a big nose. We follow it."

What this is

• An open DBT data product — every model, metric, and transformation is public, forkable, and reviewed.
• A public dataset for research, journalism, and trend analysis.
• A place to learn data analysis, analytical engineering, and data engineering while contributing to collective knowledge.

Quick start

A dbt project on Google BigQuery for Nostr event data. You need:

1. Google BigQuery credentials with read/write permissions (profiles.yml + key — see replit.md / setup_gcp_credentials.sh).
2. A Nostr event stream table (preferably partitioned) as the source for the models — e.g. Nostr.events (ingestion-time partitioned).

Project structure

• models/staging/ — staging models: clean and standardize raw data
• models/intermediate/ — intermediate transforms
• models/marts/ — business-logic / analytics marts
  • fct_network_daily_stats — the network pulse by kind
  • fct_user_activity_metrics — DAU/WAU, power users, weighted activity
  • fct_content_engagement — trending content by engagement score
  • fct_follows / fct_unfollows_daily — the social graph
  • fct_top_engaged_notes — what's catching fire
• seeds/ — sample CSV data files

Contributing

Everyone is welcome. See CONTRIBUTING.md — feature branches, tests, receipts, and no direct-main pushes.

Data ethics

All source data is public relay content, but we guard it: no PII enrichment, deletions honored where possible, no surveillance/doxxing tooling, and the same data & rules for everyone. Full policy: CONTRIBUTING.md.

License

MIT

---

© 2026 bigNOSTRals contributors · nostr